Can ambari manage an exists hadoop cluster an update it to hdp?And how?

[#1]

I used to install hadoop apache and use user defined file path.

Now I want to use ambari and hdp version,but I have a lot of data ,about tens of TB data.

Is there a way to use ambari manage the exists hadoop cluster?

And then I want to upgrade to ambari and hdp version,is there a way not stop the exists cluster for a too long time?

[#2 · Accepted solution]

Hi @darkz yu, you need to perform the so-called Takeover by Ambari, check this for your options. The easiest way to do it, provided you don't have a lot of data is a variant of option 1 from that post, meaning to export all important data from your current cluster, setup a new cluster using Ambari, and then import your data into new cluster. Note that you will need to export/import hdfs files, Hive tables, and HBase tables separately. If you want to do it keeping your data "in place" then you can consider options 2 and 3: takeover using Ambari REST API, or by using a dummy cluster. Both are more complicated than option 1.

[#3]

@darkz yu - this is an Ambari takeover situation. However, there are some manual steps involved and there are some tricky situations, so we recommend a professional service engagement to help make this successful. That will include both having Ambari take over the existing cluster as well as perform the upgrade.